WO2020093939A1

WO2020093939A1 - Positioning method, device, and electronic apparatus

Info

Publication number: WO2020093939A1
Application number: PCT/CN2019/114961
Authority: WO
Inventors: 陈成; 周帅; 徐强
Original assignee: 阿里巴巴集团控股有限公司
Priority date: 2018-11-09
Filing date: 2019-11-01
Publication date: 2020-05-14
Also published as: US20210263167A1; CN111174777A

Abstract

A positioning method, a device, and an electronic apparatus. The method comprises: acquiring, on the basis of a GNSS position of a vehicle, reference positioning data of the road surroundings of the vehicle (S320); extracting, from laser point cloud data of the road surroundings output by a laser sensor of the vehicle, positioning data awaiting matching of the road surroundings (S330, S340), the reference positioning data and the positioning data awaiting matching both comprising laser point data of key points of easily identifiable road objects having stable properties on the road around the vehicle and/or on two sides of the road; performing, on the basis of one or more sampling positions corresponding to the vehicle at a previous time instance, forward simulation of a motion state of the vehicle, and acquiring one or more sampling positions corresponding to the vehicle at the present time instance (S350); converting, on the basis of the one or more sampling positions corresponding to the vehicle at the present time instance, the positioning data awaiting matching into a coordinate system corresponding to the reference positioning data, matching the same against the reference positioning data, and obtaining the probability of the vehicle being located at each sampling position (S360, S370); and obtaining, on the basis of the probability of the vehicle being located at each sampling position, a high-precision position of the vehicle (S380). The provided positioning method determines a high-precision position of a vehicle in a fast and accurate manner.

Description

Positioning method, device and electronic equipment

This application requires the priority of the Chinese patent application filed on November 9, 2018 with the application number 201811333871.X and the invention titled "Positioning Method, Device and Electronic Equipment", the entire contents of which are incorporated by reference in this application.

Technical field

This application relates to the field of positioning technology, in particular to a positioning method, device and electronic equipment.

Background technique

The traditional vehicle positioning method is generally based on the Global Navigation Satellite System (GNSS) receiver mounted on the vehicle to obtain the real-time position of the vehicle, and the position accuracy is generally at the meter level. After the generation of high-precision maps, a positioning method based on high-precision maps emerged, that is, real-time acquisition of environmental information around the vehicle during the driving process of the vehicle, and by matching this environmental information with pre-made high-precision positioning data, the High-precision positioning results, the positioning accuracy of high-precision positioning results is generally in the centimeter level, which can meet the needs of autonomous driving. During the research on the existing high-precision map-based positioning method, the inventor found that how to quickly and accurately determine the high-precision positioning location of the vehicle is an urgent problem to be solved.

Summary of the invention

The invention provides a positioning method, device and electronic equipment, which can quickly and accurately determine the high-precision positioning position of a vehicle.

To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

In the first aspect, a positioning method is provided, including:

Obtain the vehicle's GNSS positioning location;

Based on the vehicle's GNSS positioning position, standard positioning data around the road where the vehicle is located is obtained from preset standard positioning data, and the standard positioning data includes keys of road objects on the road and / or both sides of the road that are easy to identify and have stable attributes Point laser point data;

Obtain the laser point cloud data from the vehicle's laser sensor around the road where the vehicle is located;

From the laser point cloud data, extract to-be-matched positioning data around the road where the vehicle is located, the to-be-matched positioning data includes lasers of key points of road objects on the road where the vehicle is located and / or on both sides of the road that are easy to identify and have stable attributes Point data

Based on more than one sampling position corresponding to the vehicle at the previous moment, forward simulating the movement state of the vehicle to obtain more than one sampling position corresponding to the vehicle at this moment;

Convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data based on more than one sampling position corresponding to the vehicle at this moment;

Match the positioning data to be matched with the standard positioning data in the transformed coordinate system, and based on the matching result, obtain the probability that the vehicle is located at each sampling position;

Based on the probability that the vehicle is located at each sampling location, the highly precise location of the vehicle is obtained.

In a second aspect, a positioning device is provided, including:

GNSS positioning position acquisition module, used to obtain the GNSS positioning position of the vehicle;

The standard positioning data acquisition module is used to obtain standard positioning data around the road on which the vehicle is located based on the GNSS positioning position of the vehicle, the standard positioning data including the road and / or both sides of the road is easy to identify And the laser point data of key points of road objects with stable attributes;

The laser point cloud data acquisition module is used to obtain laser point cloud data around the road where the vehicle is output by the laser sensor of the vehicle;

A positioning data extraction module to be matched is used to extract positioning data to be matched around the road where the vehicle is located from the laser point cloud data, the positioning data to be matched includes the road on which the vehicle is located and / or both sides of the road are easy to identify and have attributes Laser point data of key points of stable road objects;

The sampling position acquisition module is used to forward simulate the movement state of the vehicle based on more than one sampling position corresponding to the vehicle at the previous moment, and acquire more than one sampling position corresponding to the vehicle at the moment;

The positioning data conversion module is used to convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data based on more than one sampling position corresponding to the vehicle at this moment;

The positioning data matching module is used to match the positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, obtain the probability that the vehicle is located at each sampling position;

The high-precision positioning module is used to obtain the high-precision positioning position of the vehicle based on the probability that the vehicle is located at each sampling position.

In a third aspect, an electronic device is provided, including:

Memory for storing programs;

A processor, coupled to the memory, is configured to execute the program, and when the program is running, execute the positioning method provided by the present invention.

The invention provides a positioning method, device and electronic equipment. Based on the acquired GNSS positioning position of the vehicle, the standard positioning data around the road where the vehicle is located is obtained from the preset standard positioning data; the vehicle output from the laser sensor of the vehicle is obtained Laser point cloud data around the road, and from the laser point cloud data, extract the positioning data to be matched around the road where the vehicle is located; the above standard positioning data and the positioning data to be matched both include the road where the vehicle is located and / or both sides Laser point data of key points of road objects that are identified and stable in attributes; based on more than one sampling position corresponding to the vehicle at the previous time, forwardly simulating the movement state of the vehicle to obtain more than one sampling position corresponding to the vehicle at this time, Based on more than one sampling position corresponding to the vehicle at this moment, the positioning data to be matched is converted into the coordinate system corresponding to the standard positioning data; the positioning data to be matched in the converted coordinate system is matched with the standard positioning data, based on the matching result, Get the probability that the vehicle is at each sampling location The probability of the vehicle is located based on the position of each sample, to obtain high-precision position location of the vehicle. Since the road object of the present invention is a road object that is easy to recognize and has stable attributes on the road and / or on both sides of the road, these road objects generally do not change due to changes in the environment or over time. Therefore, the road objects and / or Or the laser point data of the key points of the road objects that are easy to identify on both sides of the road and have stable attributes can be used as high-precision positioning matching objects to ensure the positioning success rate and accuracy. At the same time, the invention only extracts laser point data of key points of road objects for matching, so the amount of data is less, the calculation amount is greatly reduced, and the positioning efficiency is improved.

The above description is only an overview of the technical solutions of this application. In order to understand the technical means of this application more clearly, it can be implemented according to the content of the specification, and to make the above and other purposes, features and advantages of this application more obvious and understandable. The specific implementation of this application is listed below.

BRIEF DESCRIPTION

By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments, and are not considered as limitations to the present application. Furthermore, the same reference numerals are used to denote the same parts throughout the drawings. In the drawings:

1a is a schematic structural diagram of a laser point cloud data collection device according to an embodiment of the present invention;

1b is a schematic diagram of a technical solution for generating positioning data according to an embodiment of the present invention;

2 is a structural diagram of a positioning system according to an embodiment of the invention;

3a is a flowchart 1 of a positioning method according to an embodiment of the present invention;

3b is a schematic diagram of a laser point cloud according to an embodiment of the invention;

4a is a flowchart 1 of a method for extracting positioning data to be matched according to an embodiment of the present invention;

4b is a scanning line diagram of an original laser point cloud according to an embodiment of the invention;

5 is a flowchart 2 of a method for extracting positioning data to be matched according to an embodiment of the present invention;

6 is a flowchart 3 of a method for extracting positioning data to be matched according to an embodiment of the present invention;

7a is a flowchart 4 of a method for extracting positioning data to be matched according to an embodiment of the present invention;

7b is an original laser point cloud image of the area on both sides of the road according to an embodiment of the present invention;

8a is a flowchart 5 of a method for extracting positioning data to be matched according to an embodiment of the present invention;

8b is a laser point cloud diagram of points of upright objects on both sides of a road according to an embodiment of the present invention;

9a is a flowchart 6 of a method for extracting positioning data to be matched according to an embodiment of the present invention;

9b is a laser point cloud diagram of ground marking points, edge points on both sides of the road, and points of upright objects on both sides of the road according to an embodiment of the present invention;

10 is a flowchart 1 of a method for acquiring a sampling position of a vehicle according to an embodiment of the present invention;

11 is a flowchart 2 of a positioning method according to an embodiment of the present invention;

12 is a flowchart 3 of a positioning method according to an embodiment of the present invention;

13 is a structural diagram 1 of a positioning device according to an embodiment of the invention;

14 is a structural diagram 1 of an extraction module of positioning data to be matched according to an embodiment of the present invention;

15 is an expanded structural diagram of a positioning data extraction module to be matched according to an embodiment of the present invention;

16 is a second structural diagram of a location data extraction module to be matched according to an embodiment of the present invention;

17 is a structural diagram 3 of a positioning data extraction module to be matched according to an embodiment of the present invention;

18 is a structural diagram 4 of a positioning data extraction module to be matched according to an embodiment of the present invention;

19 is a structural diagram 5 of a positioning data extraction module to be matched according to an embodiment of the present invention;

20 is a structural diagram of a sampling position acquisition module according to an embodiment of the present invention;

21 is a structural diagram of a positioning data matching module according to an embodiment of the present invention;

22 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

detailed description

Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

In order to achieve high-precision positioning of automobiles, it is necessary to generate positioning data for high-precision positioning scenarios. The positioning data needs to meet the following requirements:

Information volume: The information volume of the positioning data should be rich enough to represent the road and surrounding environment of the car as realistic as possible;

Data volume: The data volume of positioning data should be as small as possible to facilitate storage and transmission;

Robustness: The positioning data is sufficiently robust to the external environment such as light, time, season, climate, road conditions, etc., and is not easily affected by changes in the external environment;

Considering the above requirements, the present invention proposes a method for generating positioning data. The method includes:

Obtain the laser point cloud data of the road and / or both sides of the road within the preset area;

Extract laser point data of key points of road and / or road objects on both sides of the road from the laser point cloud data. The road objects are road objects on the road and / or both sides of the road that are easy to identify and have stable attributes;

The extracted laser point data of key points is stored as road positioning data.

In practical applications, the road objects that are easy to recognize on the road or on both sides of the road and have stable attributes may be ground markings, road edges, and upright objects on the road side.

Among them, the ground markings can be any markings painted on the road surface, such as lane lines, driving direction arrows, crosswalks, etc .; the road edges can be composed of curbstones, guardrails, green belts, etc .; the upright objects on the road side refer to the road Upright objects on both sides, for example, poles on both sides of the road (support poles for street signs, lights, traffic lights, etc.), tree trunks, walls, etc.

Due to ground markings, road edges, and upright objects on the side of the road, these road objects are not easily affected by external environments such as light, time, season, climate, and road conditions. Positioning is to match the environmental information obtained in real time during the driving process of the vehicle with positioning data Therefore, the position of the vehicle is determined. Therefore, extracting laser point data of key points of road objects that are easy to recognize and stable on the road and / or both sides of the road as positioning data can ensure the positioning success rate. At the same time, the invention only extracts laser point data of key points, so the amount of data is small, which is convenient for storage and transmission.

As shown in FIG. 1a, it is a schematic structural diagram of a laser point cloud data collection device in an embodiment of the present invention, including: a collection vehicle body 11, a wheel 12 with a wheel counter installed, and an integrated inertial measurement unit (IMU) A combined positioning system 13 with a global navigation satellite system and a lidar 14 for collecting laser point clouds. The structure of the device shown in FIG. 1a can collect laser point cloud data of all objects on the road that the vehicle travels and on both sides of the road. By processing the collected laser point cloud data through the technical solution for generating positioning data shown in FIG. 1b, positioning data with a small data amount and a high positioning success rate can be obtained.

As shown in FIG. 1b, the above technical solution for generating positioning data includes the following technical features:

S110: Acquire laser point cloud data of the road. The laser point cloud data includes the laser point data of the road and / or both sides of the road within a preset area.

S120: Divide the laser point cloud data into road surface laser point cloud data and road side laser point cloud data.

The laser point cloud data is divided into the laser point cloud data located on the road surface, the left side of the road, and the right side of the road. The specific division process can be based on the sudden ground points corresponding to the scanning lines of the laser points obtained by the lidar scan. Mutation points to distinguish the boundary position of the road surface and the laser point cloud on both sides of the road. It can be understood that if the laser point cloud data acquired in step S110 includes both the road and the laser point data on both sides of the road within the preset area, then in step S120, the road surface laser point cloud data and the road side laser point cloud can be obtained Data, if the laser point cloud data acquired in step S110 includes only one of the road and the laser point data on both sides of the road within the preset area, then step S120 can obtain road surface laser point cloud data or road side laser points Cloud data.

S130, fitting the road surface. According to the road surface laser point cloud data, a random sampling consensus algorithm (RANdom SAmple Consensus, RANSAC) is used for plane fitting to obtain the road surface.

S140, based on the fitted road surface, the height value of the road surface laser point cloud data and / or the road side laser point cloud data is corrected to a height value relative to the road surface.

Of course, if the height coordinate Z value of each laser point in the input laser point cloud data is already a Z value relative to the road surface, steps 130-140 may be omitted. Then, from the road surface laser point cloud data and / or road side laser point cloud data, correspondingly extract the laser point data of the key points of the road and / or road objects on both sides of the road.

Among them, the laser point data for extracting key points of roads and / or road objects on both sides of the road include:

S150: Extract laser point data of key points marked on the ground. Extract the laser point data of the key points of the ground mark from the road point laser point cloud data.

S160: Extract laser point data of key points on the road edge. Extract the laser point data of key points on the road edge from the road side laser point cloud data.

S170: Extract laser point data of key points of upright objects on the road side. From the laser point cloud data on the road side, the laser point data of the key points of the upright objects on the road side are extracted.

S180: Store the laser point data of the extracted key points as road positioning data.

The laser point cloud data of the ground mark key points extracted from the laser point cloud data, the road edge key points, and the key points of upright objects on the side of the road are stored as positioning data. The positioning data stored in the present invention may include any one, any two, or three types of laser point cloud data of ground marking key points, road edge key points, and key points of upright objects located on the side of the road.

Based on the positioning data generated by the above positioning data generation method, the present invention provides a positioning method, which includes:

Obtain the vehicle's GNSS positioning location;

Based on the vehicle's GNSS positioning location, the standard positioning data around the road where the vehicle is located is obtained from the preset standard positioning data. The standard positioning data includes key points of road objects on the road and / or both sides of the road that are easy to identify and have stable attributes Laser point data;

From the laser point cloud data, extract the positioning data to be matched around the road where the vehicle is located. The positioning data to be matched includes the laser point data of the key points of the road object on the road where the vehicle is located and / or both sides of the road are easy to identify and stable;

Based on more than one sampling position corresponding to the vehicle at this moment, convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data;

In the positioning method provided by the present invention, the positioning data obtained by processing the technical data generated by the positioning data shown in FIG. 1b can be used as the standard for the positioning data obtained by the laser point cloud data collected by the professional laser point cloud data collection device shown in FIG. 1a. Positioning data. The positioning data to be matched is extracted from the laser point cloud data around the road where the vehicle is located, which is output by the excitation sensor of the vehicle to be located. Both types of positioning data include laser point data of key points of road objects that are easy to identify and have stable attributes on the road and / or on both sides of the road. The difference is that standard positioning data has been clearly positioned in a GNSS coordinate system, for example. It is used as the positioning standard value; the positioning data to be matched is only the accurate positioning data output by the laser sensor of the vehicle to be positioned relative to the laser sensor coordinate system. Before matching the two to obtain the high-precision positioning position of the vehicle, Determine the rough position of the vehicle (under the GNSS coordinate system) to convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data.

In the present invention, by acquiring the GNSS positioning position of the vehicle, standard positioning data around the road where the vehicle is located can be obtained from the preset standard positioning data; based on more than one sampling position corresponding to the vehicle at the previous moment, the vehicle is simulated forward To obtain more than one sampling position corresponding to the vehicle at this moment, and then based on these sampling positions, convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data; convert the positioning data to be matched with the converted coordinate system to The standard positioning data is matched, and based on the matching result, the probability that the vehicle is located at each sampling position is obtained. Finally, based on the probability that the vehicle is located at each sampling position, the high-precision positioning position of the vehicle is obtained.

As shown in FIG. 2, it is a structural diagram of a positioning system provided by the present invention, including: a laser point cloud data collection device 210, a standard positioning database 220 and a positioning device 230; wherein:

The laser point cloud data collection device 210 may be, but not limited to, a device structure as shown in FIG. 1a (the corresponding collection vehicle body 11 may refer to a vehicle to be positioned) for collecting laser point cloud data on roads and both sides of the road, and Locate the GNSS location of the vehicle.

The standard positioning database 220 stores standard positioning data for road positioning.

The positioning device 230 is used to extract the positioning data to be matched around the road where the vehicle is located from the laser point cloud data collected by the laser point cloud data collection device 210, and at the same time obtain the GNSS positioning position of the vehicle; Obtain standard positioning data around the road where the vehicle is located in the standard positioning data of the vehicle; based on more than one sampling position corresponding to the vehicle at the previous time, forward simulate the movement state of the vehicle to obtain more than one sampling corresponding to the vehicle at this time Position, and based on more than one sampling position corresponding to the vehicle at this moment, convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data; match the positioning data to be matched with the standard positioning data in the converted coordinate system, based on the matching As a result, the probability that the vehicle is located at each sampling position is obtained, and based on the probability that the vehicle is located at each sampling position, the high-precision positioning position of the vehicle is obtained.

Wherein, the standard positioning data and the positioning data to be matched include laser point data of key points of road objects that are easy to identify and have stable attributes on the road and / or on both sides of the road. For example, the road object may be at least one road object among ground markings on the road, road edges, and upright objects located on the side of the road. These road objects generally do not change due to changes in the environment or over time. These roads The laser point data of the key points of the object is used as the positioning data of the road, which can ensure the positioning success rate. At the same time, the present invention only extracts the laser point data of the key points of the road object to match, so the data amount is less, and the calculation amount is greatly reduced To improve positioning efficiency. The technical solutions of the present application are further described below through multiple embodiments.

Example one

Based on the above positioning scheme idea, as shown in FIG. 3a, it is a flowchart 1 of a positioning method shown in an embodiment of the present invention. The method may be executed by the positioning device 230 shown in FIG. As shown in Figure 3a, the positioning method includes the following steps:

S310: Obtain the GNSS positioning position of the vehicle.

By setting the GNSS system on the vehicle that needs to be located, the vehicle can be located in real time and the GNSS positioning position of the vehicle can be obtained.

S320, based on the GNSS positioning position of the vehicle, obtain standard positioning data around the road where the vehicle is located from the preset standard positioning data, and the standard positioning data includes the key of the road object on the road and / or both sides of the road that are easy to identify and have stable attributes Point laser point data.

After obtaining the GNSS positioning location of the vehicle, you can roughly determine the geographic location of the vehicle. Obtain the standard positioning data around the road where the vehicle is located from the standard positioning database according to the GNSS positioning position of the vehicle. These standard positioning data can be obtained through the technical solution as shown in FIG. 1b, which includes easy identification on the road and / or both sides of the road and Laser point data of key points of road objects with stable attributes. Each laser point data corresponds to a clear position coordinate in the GNSS system coordinate system.

These road objects may include ground markings, road edges, and upright objects on the side of the road.

S330: Acquire laser point cloud data around the road where the vehicle is located and output by the laser sensor of the vehicle.

It should be noted that the laser point cloud at a far position (of tens of meters to hundreds of meters) can usually be scanned by lidar scanning. The laser point cloud farther away from the collection point has lower accuracy, and is usually not The laser spot at the road location. Therefore, when collecting laser point cloud data, the vehicle body can be the center, and the laser point cloud that is far away from the collected vehicle body can be directly filtered out. The filtering is only limited in scope to reduce the amount of redundant laser point cloud data. The laser point cloud data at this time does not distinguish between ground and non-ground.

When actually acquiring the laser point cloud data around the road where the vehicle is located, the laser point cloud around the road where the vehicle is currently at the current moment can be periodically obtained from the laser point cloud data output by the vehicle's laser sensor in accordance with the processing cycle for generating high-precision positioning positions data. As shown in Figure 3b, it is a laser point cloud map corresponding to a certain point in the laser point cloud data located in the preset area around the road where the vehicle is located. Each laser point data in the figure includes x, y, z three-dimensional coordinates (located in the laser Sensor coordinate system), the lightness and darkness of each laser spot characterize the reflectivity of the laser spot, where the reflectivity of the laser spot on the road surface in the middle area is higher than the reflectivity of the laser spot on both sides of the road.

S340. Extract the positioning data to be matched around the road where the vehicle is located from the laser point cloud data. The positioning data to be matched includes the laser point data of the key points of the road object on the road where the vehicle is located and / or on both sides of the road that are easy to identify and have stable attributes .

In order to ensure the consistency of the positioning data to be matched and the standard positioning data on the road object, after acquiring the laser point cloud data around the road where the vehicle is located and output by the laser sensor of the vehicle, it is necessary to extract the vehicle including the vehicle from these laser point cloud data Laser point data of key points of road objects that are easy to identify and have stable attributes on the road and / or on both sides of the road.

Among them, road objects may include but are not limited to: ground markings, road edges, and upright objects located on the side of the road. Correspondingly, the laser point data of the key points of the road objects may be the laser point data extracted from the laser point cloud data of these road objects that can best reflect the key points of the morphological characteristics of the road objects.

Specifically, the reflectivity of the laser point can be used in combination with the three-dimensional coordinate value (especially the height value of the laser point), and the road surface and / or both sides of the road can be extracted from the laser point cloud data within a predetermined area Laser point data of marks, key points of road edges and upright objects located on the side of the road.

Specifically, a method for forming standard positioning data may be used to obtain positioning data to be matched from the laser point cloud data around the road where the vehicle is output from the laser sensor of the vehicle to be located. In this way, if the positioning data to be matched and the standard positioning data are laser point data for key points of road objects in the same geographic area, the similarity of the distribution of the two positioning data should be high. The only difference is that the standard positioning data is clearly positioned with accurate GNSS positioning coordinates, while the positioning data to be matched is only for the laser sensor coordinate system, and the positioning for the GNSS positioning system is relatively rough (mainly because of the user ’s The accuracy of the GNSS positioning system used is not high), it is necessary to rely on the matching results of the positioning data to be matched and the standard positioning data to finally determine the high-precision positioning position of the vehicle.

S350, based on more than one sampling position corresponding to the vehicle at the previous moment, forwardly simulating the movement state of the vehicle, and acquiring more than one sampling position corresponding to the vehicle at this moment.

When periodically matching the positioning data to be matched with the standard positioning data of the vehicle, more than one sampling position may be set for the current vehicle in each period, and these sampling positions represent the positions where the vehicle may appear in the corresponding period.

Specifically, when determining more than one sampling position of the vehicle at the current moment, it is possible to move the more than one sampling position corresponding to the vehicle at the previous moment to the current moment by forwardly simulating the movement state of the vehicle, thereby obtaining the current moment More than one sampling location. Wherein, the sampling position of the vehicle at the initial time is more than one position point selected from the positioning area where the GNSS positioning position of the vehicle is located at the initial time.

In the actual application scenario, based on more than one sampling position corresponding to the vehicle at the previous moment, combined with vehicle kinematics, IMU measurement data, wheel number meter and other devices to simulate the movement of the vehicle, the current possible sampling position of the vehicle is obtained. If the current time is the initial time to obtain the sampling position, sampling can be directly performed in the positioning area where the positioning position output by the GNSS system is located to obtain more than one possible sampling position of the vehicle at the current time.

Among them, forward simulation of vehicle motion is based on vehicle kinematics equations. For example, the following gives an example of forward simulation:

The model in equation (1) assumes that the vehicle's turning angular rate and movement speed are constant in a short time, where

Respectively represent the x, y coordinates of the vehicle in the plane coordinate system at time _k , υ _k ,

Respectively represent the speed, heading angle and heading angle change rate of the vehicle at time k, Δt represents the time difference from time k to time k + 1, and v _k represents the model noise at time k

In this way, by simulating the vehicle motion state forward, more than one sampling position of the vehicle at each time (excluding the initial time) can be obtained.

S360, based on more than one sampling position corresponding to the vehicle at this moment, converting the positioning data to be matched into the coordinate system corresponding to the standard positioning data.

After determining the position coordinates of the sampling position of the vehicle in the coordinate system corresponding to the standard positioning data, the positioning data to be matched corresponding to the same time can be converted into the coordinate system corresponding to the standard positioning data to obtain the corresponding position coordinate values.

S370: Match the positioning data to be matched with the standard positioning data in the converted coordinate system, and obtain the probability that the vehicle is located at each sampling position based on the matching result.

For each sampling position, there will be a set of positioning data to be matched in the converted coordinate system. Match these positioning data to the standard positioning data. Based on the matching result, the positioning data to be matched in the converted coordinate system and the standard positioning data The matching degree of the laser spot data can determine the probability of the vehicle appearing at the sampling position.

S380, based on the probability that the vehicle is located at each sampling position, obtain the high-precision positioning position of the vehicle.

For example, the sampling position with the highest probability of the probability that the vehicle is located at each sampling position may be determined as the high-precision positioning position of the vehicle;

or,

Taking the probability that the vehicle is located at each sampling position as a weight, the corresponding sampling positions are weighted, and the obtained weighted position is determined as the high-precision positioning position of the vehicle.

Example 2

As shown in FIG. 4a, it is a flowchart 1 of a method for extracting positioning data to be matched according to an embodiment of the present invention. This embodiment can be used as a preferred embodiment for extracting positioning data to be matched around the road where the vehicle is located from the laser point cloud data in the method shown in FIG. 3a. As shown in FIG. 4a, the above step S340 may specifically include the following steps:

S410. Divide the laser point cloud data into road surface laser point cloud data and / or road side laser point cloud data.

Specifically, the laser point cloud data around the road where the vehicle is acquired in step S330 is divided into road surface laser point cloud data and / or road side according to the change characteristics of the position and height value of the three-dimensional coordinate value (Z value in three-dimensional coordinates) (Including the left and right side of the road) Laser point cloud data.

In the actual application scenario, find the abrupt point of the laser point cloud from each lidar scan line. For example, in the laser point cloud data in a grid, when the height difference between the highest point and the lowest point in the laser point cloud on the same scan line is greater than a certain threshold, such as 0.1m, it is considered that there is The point of high abrupt change in the laser point cloud on this scan line. For another example, in the laser point cloud data in two adjacent grids, when the height difference between the highest point and the lowest point in the laser point cloud on the same scan line is greater than a certain threshold, it may also be considered that there are The abrupt point of the laser point cloud on the scan line. By extending from the middle position of the scan line corresponding to the laser point cloud data around the road where the vehicle is located to both sides, the high point of abrupt change in the laser point cloud on the scan line can be identified, and the laser point cloud data on a scan line can be further It is divided into road surface laser point cloud data and road side laser point cloud data.

It should be noted that, for each scan line, the scan line is approximately perpendicular to the driving direction of the vehicle, and expands from the center of the scan line to the left and right sides, respectively, to find the height mutation point in the laser point cloud on the left and right scan lines. One scan line is divided into road surface laser point cloud data and road side laser point cloud data. The same operation is performed on multiple scan lines, so that the laser point cloud data around the road where the vehicle is located can be divided into road surface laser point cloud data and road side laser point cloud data.

As shown in Figure 4b, from left to right, you can see that each scan line is approximately a circular arc. The middle point of the scan line is the position of the laser point that the vehicle has / will be passing by. The laser point must be the road surface. On the laser spot. Expand from the middle of the scan line to both sides. If the height change of two adjacent laser points is greater than the specified height threshold, the position of the laser point is considered to reach the edge of the road, stop expansion, and from the position of the road edge to the scan line Laser point cloud for segmentation. Divide all the scanning lines of the laser point cloud corresponding to the laser point cloud data around the road where the vehicle is located to obtain the laser point cloud data of the three regions shown in Figure 4b, from left to right are the laser points on the left side of the road Cloud data, laser point cloud data on the road surface and laser point cloud data on the right side of the road.

S420: Extract laser point data of key points of road and / or road objects on both sides of the road from the road surface laser point cloud data and / or road side laser point cloud data.

After the area division of the laser point cloud data is completed, the laser point data of the key points of the road object can be extracted from the laser point cloud data corresponding to the different areas, such as the laser points of the key points of the ground markings extracted from the road surface laser point cloud data Data, laser point data of key points of road edges and upright objects are extracted from the road side laser point cloud data.

In addition, as shown in FIG. 5, before step S420 is performed, the following steps may also be performed, so as to correct the height of the laser point cloud data relative to the fitted road surface.

S510. Fit the road surface of the road according to the laser point cloud data of the road surface.

For example, a random sampling consensus algorithm can be used to perform plane fitting on the road surface laser point cloud data to obtain the road surface of the road.

Specifically, the RANSAC plane fitting algorithm can be used to fit the road surface laser point cloud data to fit the ground plane, and the part of the ground plane located in the road area is the road surface. The specific fitting steps are as follows:

a) Randomly extract 3 data points P ₁ , P ₂ , P ₃ from the road surface laser point cloud data;

b) Generate a plane from these 3 data points, calculate the distance of all road surface laser point data from the plane, and count the number of laser points within a certain distance (such as 5cm).

c) Repeat the above steps several times, and the plane consisting of three points with the largest amount of road surface laser point data within a certain range from the plane consisting of three points is determined as the ground plane. The part of the ground plane located in the road area is the road surface.

S520, based on the fitted road surface, correct the height values of the road surface laser point cloud data and the road side laser point cloud data to a height value relative to the road surface.

For example, if the height value corresponding to the road surface is set to height 0, the height values of the road surface laser point cloud data and road side laser point cloud data can be corrected to the distance from the corresponding laser point to the road surface.

In addition, it should be added that if the Z value in the input laser point cloud data is already the Z value relative to the road surface, no additional steps S510 to S520 need to be performed.

This embodiment is based on the embodiment shown in FIG. 3a. Further, by dividing the laser point cloud data into road surface laser point cloud data and / or road side laser point cloud data; then, from the road surface laser point cloud data and / Or the road side laser point cloud data, extract the laser point data of the key points of the road and / or road objects on both sides of the road, so as to realize the convenient and rapid acquisition of the laser point data of the key points of the road object, that is, to be matched and positioned data.

In addition, before extracting the laser point data of the key points of the road and / or road objects on both sides of the road, by fitting the road surface of the road based on the road surface laser point cloud data; and based on the fitted road surface, the The height value of the road surface laser point cloud data and the road side laser point cloud data is corrected to the height value relative to the road surface, thereby ensuring the accuracy of the height position of the laser point cloud data.

Example Three

As shown in FIG. 6, it is a flowchart 3 of a method for extracting positioning data to be matched according to an embodiment of the present invention. The difference between this embodiment and the method shown in FIG. 4a is that, in this embodiment, when the road object is a ground mark on the road, the laser point data of the key points of the road object of the road is extracted from the road surface laser point cloud data A preferred embodiment. As shown in Figure 6, the following steps can be performed initially at the method:

S411: Divide the laser point cloud data into road surface laser point cloud data.

This step may be a specific way of dividing the laser point cloud data in step S410.

S610: Divide the road surface laser point cloud data into a grid according to a preset grid size.

The preset grid may be a two-dimensional grid set on a horizontal plane. According to the projection relationship between the road surface laser point cloud data and the grid, all road surface laser point cloud data may be divided into different grids.

S620: If the road surface laser point cloud data in a grid includes the laser light of the ground mark, then based on the laser light of the ground mark in the grid, obtain the laser point data of a key point of the ground mark.

In the road surface laser point cloud data, the difference between the reflectivity of the laser point cloud marked on the ground and the laser point cloud not marked on the ground is obvious. Generally, the ground area with ground marks usually corresponds to the lane line, arrow, crosswalk, etc. on the road, so compared with the laser point cloud of other non-ground mark ground areas, the reflectivity of the laser point cloud in this part of the ground area Too large. Based on this feature, the laser point data of ground markings can be extracted from the road point laser point cloud data.

For example, you can count the number of laser points in each grid, the mean and variance of the reflectivity of the laser points; then, determine the laser point data that meets the number threshold, mean threshold, and variance threshold specified in the preset conditions Laser point data marked for the ground.

Specifically, the preset conditions can be set according to the characteristics of the laser points in the grid containing ground marks according to pre-learning or experience, and the preset conditions can specify the number threshold and reflectivity of the laser points in the grid containing ground marks Indicators such as the mean threshold value and the variance threshold value, when the laser point in the grid to be processed meets the requirements of the preset conditions, it is determined that the corresponding laser point is the laser point of the ground mark. For example, if the number of laser points in the grid, the mean and variance of the reflectance of the laser points all reach the specified preset, and the reflectance of the current laser point is greater than the specified value exceeding the average, it can be determined that the laser point is a laser point marked on the ground.

If the road surface laser point cloud data in a grid contains the laser point data of the ground mark, the laser point data of a key point of the ground mark can be obtained based on the laser point data of the ground mark in the grid. For example, when there are multiple laser point data of ground marks in a grid, the laser point data of a key point of the ground mark can be obtained based on the average value of the multiple laser point data. For example, the average value of the coordinates (xyz) in the laser point data is calculated, and then the obtained coordinate average value is used as the coordinate of the laser point data of the key point of the ground mark.

This embodiment is based on the embodiment shown in FIG. 4a, and further, by determining the road object as a ground mark on the road and dividing the road surface laser point cloud data into a grid according to a preset grid size; When it is determined that the road surface laser point cloud data in a grid contains the laser light of the ground mark, then based on the laser light of the ground mark in the grid, the laser point data of a key point of the ground mark is obtained to achieve Obtain the laser point data of the key points of ground marking conveniently and quickly.

Example 4

As shown in FIG. 7a, it is a flowchart 4 of a method for extracting positioning data to be matched according to an embodiment of the present invention. The difference between this embodiment and the method shown in FIG. 4a is that, in this embodiment, when the road object is a road edge, the laser point data of the key points of the road objects on both sides of the road are extracted from the road side laser point cloud data A preferred embodiment. As shown in Figure 7a, the following steps can be performed initially at the method:

S412: Divide the laser point cloud data into road side laser point cloud data.

S710: Divide the laser point cloud data on the road side into a grid according to a preset grid size.

The preset grid may be a two-dimensional grid set on a horizontal plane. According to the projection relationship between the road-side laser point cloud data and the grid, all road-side laser point cloud data may be divided into different grids.

S720: If the road-side laser point cloud data in a grid contains laser-point data on the road edge, the laser-point data on the road edge is sorted according to the height values in the laser point data in ascending order.

Among them, the laser point data near the area contiguous with the road in the laser point cloud data on the left side of the road can be recorded as the laser point data on the left edge of the road, and the laser near the area contiguous with the road in the laser point cloud data on the right side of the road Point data is recorded as laser point data on the right edge of the road.

In actual application scenarios, the laser point data near the area closest to the collected vehicle driving trajectory can also be obtained from the left side of the road and the right side of the road as the laser point data of the road edge.

As shown in FIG. 7b, for the laser point cloud data on both sides of the road, the laser point data near the boundary point closest to the road can be extracted from the areas on both sides as the laser point data on the road edge.

When the road side laser point cloud data in a grid contains the laser edge data of the road edge, the laser point data of these road edges are sorted in order of the height value in the laser point data from small to large. In the sorting process, the laser point data of the road edges can be sorted separately according to their corresponding left and right road edges, or they can be put together for unified sorting.

S730: If the difference between the height values of the two adjacent laser points after the sorting is greater than the preset difference threshold, delete the laser points after the sorting and the subsequent laser points from the two adjacent laser points.

In an actual application scenario, if the difference between the height values of two adjacent laser points after the above sorting is greater than the preset difference threshold, it is likely that the location of the two laser points is between the road and the area on both sides of the road The boundary between the laser point in the rear and the laser point in the subsequent order is likely to correspond to the edge position of the sudden change of the height of the curb, the guardrail, the green belt on both sides of the road, or the dangling point. In this case, you can delete the laser points that are sorted in the next two laser points and the laser points that follow, that is, the laser point data of the part of the road edge farther from the road, and keep the two laser points that are sorted in the next The previous and previous laser points, that is, the laser point data on the edge of the part of the road closer to the road, ensure the quality of the data to be processed later, while reducing the amount of data to be processed.

S740. Acquire laser point data of a key point on the road edge based on the laser point data of the road edge retained in the grid.

For example, you can select one laser point data from the laser point data of the road edge retained in the grid as the laser point data of the key point, or when there are multiple laser point data of the road edge retained in a grid , The laser point data of a key point on the road edge can be obtained based on the average value of the multiple laser point data. For example, calculate the average of the coordinates (xyz) in these laser point data, and then use the obtained coordinate average as the coordinates of the laser point data of the key points on the road edge.

This embodiment is based on the embodiment shown in FIG. 4a, and further, by determining the road object as a road edge, and dividing the road side laser point cloud data into a grid according to a preset grid size; when determining a When the road side laser point cloud data in the grid contains the laser edge data of the road edge, the laser edge data of the road edge is sorted according to the order of the height values in the laser point data; if the two are adjacent after sorting If the difference between the height values of the laser points is greater than the preset difference threshold, the laser points that are ranked after and the laser points that follow are deleted from the two adjacent laser points; finally, based on the road edge reserved in the grid Laser point data, to obtain laser point data of a key point on the road edge, so as to achieve convenient and fast access to the laser point data of the key point on the road edge.

Example 5

As shown in FIG. 8a, it is a flowchart 5 of a method for extracting positioning data to be matched according to an embodiment of the present invention. The difference between this embodiment and the method shown in FIG. 4a is that, in this embodiment, when the road object is an upright object located on the side of the road, the key points of the road objects on both sides of the road are extracted from the road side laser point cloud data A preferred embodiment of laser spot data. As shown in Figure 8a, the following steps can be performed initially at the method:

S412: Divide the laser point cloud data into road side laser point cloud data.

S810. If the road-side laser point cloud data in a grid includes laser-point data of road-side upright objects, then the laser-point data of the road-side upright objects in order from the height value in the laser point data to small Sort.

Specifically, the laser point data with the height within the preset height range can be extracted from the laser point cloud data on the left side and right side of the road as the laser point data of the upright object on the road side.

For example, a height threshold (such as greater than 0.5m and less than 2.5m) can be preset to clear the laser point cloud data outside the height threshold on both sides of the road, and the remaining laser point cloud data is the selected location Laser point data of upright objects on the road side.

As shown in FIG. 8b, it is laser point cloud data of upright objects on both sides of the road extracted from both sides of the road.

When the road-side laser point cloud data in a grid contains the laser point data of road-side upright objects, the laser point data of these road-side upright objects is sorted in order from the height value in the laser point data. In the sorting process, the laser point data of upright objects on the road side can be sorted separately according to their corresponding upright objects on the left side of the road and upright objects on the right side of the road, or they can be put together and sorted together.

S820: If the difference in height between the two adjacent laser spots after sorting is greater than the preset difference threshold, delete the laser spots that are ranked after and in the two adjacent laser spots.

In actual application scenarios, if the difference between the height values of two adjacent laser points after the above sorting is greater than the preset difference threshold, it is likely that the positions of the two laser points are two upright objects in the road side area The edge of the laser beam, and the laser spot after it and the laser spot after it are likely to correspond to the position where the height of the pole (street sign, lighting lamp, traffic light, etc.) changes suddenly, or the height of the tree, wall, etc. point. In this case, you can delete the laser points in the next two laser spots and the laser spots in the subsequent order, and keep the laser spots in the two adjacent laser spots in order before and before, to ensure the quality of the subsequent processing data , While reducing the amount of data to be processed.

S830: Determine whether the lowest height value in the laser point data of the retained upright object is less than the preset first height threshold, and whether the highest height value is greater than the preset second height threshold, and if so, based on the Laser point data of an upright object, to obtain laser point data of a key point of the upright object.

The first height threshold is smaller than the second height threshold.

This step is to further determine whether these data satisfy the corresponding upright object and still meet a certain height range in the laser point data of the upright object retained. If satisfied, based on the laser point data of the upright object retained in the grid, the laser point data of a key point of the upright object is obtained.

For example, one of the laser point data of the upright objects retained in the grid can be selected as the laser point data of the key point, or when there are multiple laser point data of the upright objects retained in a grid , The laser point data of a key point of an upright object can be obtained based on the average value of the multiple laser point data. For example, calculate the average of the coordinates (xyz) in these laser point data, and then use the obtained coordinate average as the coordinates of the laser point data of the key point of the upright object.

This embodiment is based on the embodiment shown in FIG. 4a, and further, by determining the road object as an upright object located on the road side, and dividing the road side laser point cloud data into grids according to a preset grid size ; When it is judged that the road-side laser point cloud data in a grid contains the laser-point data of the road-side upright objects, the laser-point data of the road-side upright objects is sorted according to the order of the height values in the laser point data from small to large ; If the difference between the height values of the two adjacent laser points after sorting is greater than the preset difference threshold, delete the laser points that are ranked after and after the laser points in the adjacent two laser points; finally, judge to keep Whether the lowest height value in the laser point data of the upright object is less than the preset first height threshold, and whether the highest height value is greater than the preset second height threshold, if so, based on the laser of the upright object retained in the grid Point data, to obtain laser point data of a key point of an upright object, so as to achieve convenient and fast acquisition of laser point data of key points of an upright object on the road side.

Example Six

As shown in FIG. 9a, it is a flowchart 6 of a method for extracting positioning data to be matched according to an embodiment of the present invention. The difference between this embodiment and the method shown in FIG. 4a is that this embodiment uses the laser point cloud data from the road surface and the road side laser point cloud data when the road object includes ground markings, road edges, and upright objects on the road side In a preferred embodiment of extracting laser point data of key points of roads and road objects on both sides of the road. As shown in Figure 9a, the following steps can be performed initially at the method:

S413: Divide the laser point cloud data into road surface laser point cloud data and road side laser point cloud data.

S910: Divide the laser point cloud data of the road surface and the laser point cloud data of the road side into a grid according to a preset grid size.

S920, if the road surface laser point cloud data in a grid contains the laser data of the ground mark, then based on the laser data of the ground mark in the grid, the laser point data of a key point of the ground mark is obtained.

S930, if the road-side laser point cloud data in a grid contains laser-point data on the road edge, the laser-point data on the road edge is sorted in descending order of the height value in the laser point data;

S940: If the difference between the height values of the two adjacent laser points after the sorting is greater than the preset difference threshold, delete the laser points after the sorting and the subsequent laser points from the two adjacent laser points.

S950. Acquire laser point data of a key point on the road edge based on the laser point data of the road edge retained in the grid.

S960, if the road-side laser point cloud data in a grid contains the laser point data of the road-side upright object, the laser point data of the road-side upright object is arranged in the order of the height value in the laser point data from small to large Sort.

S970: If the height difference between the two adjacent laser spots after sorting is greater than the preset difference threshold, delete the laser spots that are ranked after and after the laser spots in the two adjacent laser spots.

S980: Determine whether the lowest height value in the laser point data of the retained upright object is less than the preset first height threshold, and whether the highest height value is greater than the preset second height threshold, and if so, based on the Laser point data of an upright object, to obtain laser point data of a key point of the upright object.

For the specific content of steps S910 to S960, please refer to the content of similar steps in FIG. 6, FIG. 7 a and FIG. 8 a, which will not be repeated here.

In actual application scenarios, in order to reduce the dangling points and stray points in the extracted road surface laser point cloud data and road side laser point cloud data, the laser point cloud data can be filtered first after the laser point cloud data is obtained The dangling points in the map, so that the filtered laser point cloud data correspond to real and effective environmental data as much as possible. For example, you can divide the road surface laser point cloud data and road side laser point cloud data into a grid, and then sort them by the height value of the laser point, filter out the dangling points in the grid, and only keep the entities that continue from the road surface upward point. This process can effectively filter out laser point data of suspended objects such as branches other than the main pole of the tree. .

In summary, as shown in FIG. 9b, it is a schematic diagram of a laser point cloud extracted from ground point marks, road edges, and key points of upright objects on the side of the road extracted from laser point cloud data around the road where the vehicle is located.

Based on the embodiment shown in FIG. 4a, this embodiment further determines the road surface laser point cloud data and the road side laser point by determining the road object as a ground mark on the road, the road edge and an upright object located on the road side Cloud data is divided into grids according to the preset grid size; and based on the laser point data of ground markings in the grid, the laser point data of road edges, and the laser point data of upright objects located on the side of the road, the corresponding roads are obtained respectively The laser point data of a key point of the object, so as to realize the convenient and rapid acquisition of the laser point data of the key points of ground markings, road edges and upright objects located on the side of the road.

Example 7

As shown in FIG. 10, it is a flowchart 1 of a method for acquiring a sampling position of a vehicle according to an embodiment of the present invention. This embodiment can be used as a preferred implementation of the method shown in FIG. 3a, based on more than one sampling position corresponding to the vehicle at the previous time, to forwardly simulate the movement state of the vehicle, and obtaining more than one sampling position corresponding to the vehicle at the current time. Program. As shown in FIG. 10, the above step S350 may specifically include the following steps:

S101. Extract the first sampling position where the probability that the vehicle is at the corresponding sampling position is greater than the probability threshold from more than one sampling position corresponding to the vehicle at the previous moment.

Specifically, in the result of calculating the probability of the vehicle at each sampling position at the previous moment, a sampling position whose probability value is greater than a preset probability threshold may be selected as the first sampling position, and these first sampling positions are relative to other sampling positions, The probability of the vehicle appearing at this sampling location will be higher.

S103 forwardly simulates the movement state of the vehicle for the first sampling position, and obtains more than one sampling position corresponding to the vehicle at this moment.

Based on the first sampling position, the sampling position of the vehicle at this moment generated by the forward simulated vehicle movement is closer to the actual position of the vehicle than other sampling positions, so that the vehicle is accurately and quickly positioned with high precision. In addition, in this embodiment, in order to maintain the number of sampling positions used at each time, as shown in FIG. 10, the following steps may also be performed after step S101 and before S103:

S102. Select multiple location points near the first sampling position as the additional first sampling position.

After selecting a partial sampling position from the one or more sampling positions of the vehicle used at the previous moment as the first sampling position for the sampling position of the vehicle at this moment, although the accuracy of the sampling position at this moment is improved, the sampling position The number of will be reduced, in order to keep the number of sampling locations of the vehicle at this moment, or maintain the first level, you can select multiple location points near the first sampling location after extracting the first sampling location each time As the additional first sampling position. Since the additional first sampling positions are located near the original first sampling positions, the accuracy of these sampling positions can still be guaranteed.

This embodiment is based on the embodiment shown in FIG. 3a, and further, by extracting the first sampling position where the probability that the vehicle is at the corresponding sampling position is greater than the probability threshold from more than one sampling position corresponding to the vehicle at the previous moment; and For the first sampling position, forwardly simulate the movement state of the vehicle to obtain more than one sampling position corresponding to the vehicle at this moment, so as to achieve convenient and rapid acquisition of multiple sampling positions of the vehicle used at this moment, and ensure the sampling position Accuracy. In addition, by selecting a plurality of position points near the first sampling position as the additional first sampling position, the number of first sampling positions can be maintained at a certain level, while ensuring the accuracy of the sampling position.

Example 8

As shown in FIG. 11, it is a flowchart 2 of a positioning method according to an embodiment of the present invention. The difference between this embodiment and any one of the positioning methods shown in FIG. 3a to FIG. 10 is that this embodiment adopts matching the positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, the vehicle is located in each position. A preferred embodiment of the probability of each sampling location. As shown in FIG. 11, taking FIG. 3a as an example, the following steps can be performed after step S360:

S111 matches the positioning data to be matched in the converted coordinate system with the laser point data of the key points of the road object of the same type in the standard positioning data, and based on the matching result, obtains the probability that the vehicle is located at each sampling position.

Specifically, when matching the positioning data to be matched with the standard positioning data in the converted coordinate system, the matching may be performed according to the type classification of the laser point data of the key points of the road object included in the positioning data. For example, the road objects are all the laser point data of the key points of the ground markings on the road, the road objects are the laser point data of the key points of the road edge, and the road objects are the key points of the upright objects on the side of the road The laser point data are matched separately to improve the accuracy of the matching results. Obtaining the probability that the vehicle is located at each sampling position based on the matching result can improve the accuracy of the probability that the vehicle is located at each sampling position.

Further, as shown in FIG. 12, it is a flowchart 3 of a positioning method according to an embodiment of the present invention. The difference between this embodiment and any one of the positioning methods shown in FIG. 3a to FIG. 10 is that this embodiment adopts matching the positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, the vehicle is located in each position. A preferred embodiment of the probability of each sampling location. As shown in FIG. 12, after step S360, the following steps may be performed:

S121 Match the laser point data of the key point in the positioning data to be matched in the conversion coordinate system with the laser point data of the closest key point in the standard positioning data, and calculate the key point and the standard in the positioning data to be matched in the conversion coordinate system The probability that the closest key point in the positioning data is the same location point.

For each possible sampling position of the vehicle, the probability of the vehicle appearing at the sampling position needs to be calculated. This probability can correspond to the laser point data of the key point in the positioning data to be matched of the converted coordinate system and the closest distance to the standard positioning data The laser point data of the key points is the matching of points at the same position.

Specifically, for the laser point cloud data of each key point in the positioning data to be matched of the conversion coordinate system, find the key point in the standard positioning data closest to the laser point cloud data space of the key point in the standard positioning data For laser point cloud data, the difference between the laser point cloud coordinates of these two key points is Δx, Δy, Δz. It is assumed that the laser point cloud data of each key point in the standard positioning data obey the mean value

The variance is

Three-dimensional normal distribution. Then the distribution probability of the laser point cloud data P of each key point of the positioning data to be matched with respect to the laser point cloud data of the closest key point in the standard positioning data is:

S122 obtains the probability that the vehicle is located at the sampling position according to the probability that the key point in the positioning data to be matched of all conversion coordinate systems corresponding to the same sampling position and the closest key point in the standard positioning data are the same position point.

For example, for a group of vehicles, the number of sampling positions is 500, and the number of laser point cloud data of the key points of the positioning data to be matched is 1000 of the sample space. If for each sampling position, the key points of the positioning data to be matched Laser point cloud data, the laser point data of the key point in the standard positioning data closest to the laser point cloud data is {P ₁ , P ₂ , ..., P _N }, N is 1000, then you can get The probability of sampling positions is:

Through the above process, the probability of the vehicle at each sampling position can be calculated.

This embodiment is based on the embodiments shown in FIGS. 3a to 10, and further, by matching the positioning data to be matched of the converted coordinate system with the laser point data of the key points of the road object of the same type in the standard positioning data, Based on the matching result, the probability that the vehicle is located at each sampling position is obtained, so that in the process of matching the laser point data of the key point, the comparison between road objects of the same type is ensured to improve the accuracy of the matching result, and thereby improve the vehicle location The accuracy of the probability of each sampling location.

In addition, by matching the laser point data of the key points in the positioning data to be matched in the converted coordinate system with the laser point data of the closest key points in the standard positioning data, the key points in the positioning data to be matched in the converted coordinate system are calculated Probability that the closest key point in the standard positioning data is the same position point; according to the same sampling position, the key point in the positioning data to be matched of all conversion coordinate systems and the closest key point in the standard positioning data are the same position point The probability of obtaining the probability that the vehicle is at the sampling position can be used to quantify the probability of the vehicle at each sampling position, while improving the accuracy of the calculation result.

In an actual application scenario, the method steps shown in FIG. 11 and FIG. 12 can be combined in one execution of the positioning method.

In addition, in the traditional laser point cloud-based positioning method, the standard positioning data used is mainly obtained by the following methods, including:

Laser point cloud 3D space thinning: The original laser point cloud has a large amount of data. By dividing the 3D space into several grids (such as 10x10x10cm), storing one laser point in each grid reduces the laser point cloud data. the amount;

Laser point cloud ground thinning: first extract the ground from the original point cloud, and then rasterize the ground point cloud in a two-dimensional space, each grid only stores the statistical information of the reflectivity of the ground point cloud;

Laser point cloud thinning on both sides of the road: first generate a road reference line, then project the laser point cloud perpendicular to the reference line to both sides of the road, only retain the laser point closest to the reference line within a certain height range, and raster the laser point After gridding, each raster is stored in the positioning layer.

But these methods have certain defects:

The first scheme has too much data, which is not conducive to storage and matching positioning; some data in the environment (such as shrubs, branches, etc.) will change with time, season, and climate, making it difficult to locate effectively.

The second scheme only retains the reflectivity of the ground laser point cloud. In the case of water or snow on the ground, it is difficult to accurately obtain the ground reflectivity and cannot be matched and positioned.

The second scheme relies on reference lines and generates too many map links; some laser point clouds (such as shrubs, branches, etc.) on both sides of the road will change with time, season, and climate, making it difficult to locate effectively; at the same time, because this scheme only stores road two Side data, if there are other vehicles on both sides of the self-driving car, it will affect the positioning result.

The acquisition process of standard positioning data used in the positioning method provided by the present invention makes up for the defects in the existing methods. Road objects that are easy to identify and have stable attributes on the road and / or on both sides of the road are used as road objects, and these roads are extracted The laser point data of the key points of the object is used as the positioning data of the road. These road objects generally do not change due to changes in the environment or over time, and positioning is to match the environmental information obtained in real time during the driving process of the vehicle with the positioning data to determine the position of the vehicle. Therefore, extract the road and / or Or the laser point data of key points of road objects that are easy to identify on both sides of the road and have stable attributes can be used as positioning data to ensure the positioning success rate. At the same time, the present invention only extracts laser point data of key points, so the data amount is small, which is convenient for storage and transmission. When high-precision positioning of the vehicle is performed later, the calculation amount can also be reduced and the positioning efficiency can be improved.

Example 9

As shown in FIG. 13, FIG. 1 is a structural diagram of a positioning device according to an embodiment of the present invention. The positioning device may be provided in the positioning system shown in FIG. 2 and used to perform the method steps shown in FIG. 3 a, which include:

The GNSS positioning position obtaining module 131 is used to obtain the GNSS positioning position of the vehicle;

The standard positioning data acquisition module 132 is used to obtain standard positioning data around the road where the vehicle is located from preset standard positioning data based on the vehicle's GNSS positioning position. The standard positioning data includes roads and / or sides of the road that are easy to identify and Laser point data of key points of road objects with stable attributes;

The laser point cloud data acquisition module 133 is used to acquire laser point cloud data around the road where the vehicle is located, which is output by the laser sensor of the vehicle;

The to-be-matched positioning data extraction module 134 is used to extract to-be-matched positioning data around the road where the vehicle is located from the laser point cloud data. Laser point data of key points of the object;

The sampling position obtaining module 135 is used to forwardly simulate the movement state of the vehicle based on more than one sampling position corresponding to the vehicle at the previous moment, and obtain more than one sampling position corresponding to the vehicle at this moment;

The positioning data conversion module 136 is used to convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data based on more than one sampling position corresponding to the vehicle at this moment;

The positioning data matching module 137 is used to match the positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, obtain the probability that the vehicle is located at each sampling position;

The high-precision positioning module 138 is used to obtain the high-precision positioning position of the vehicle based on the probability that the vehicle is located at each sampling position.

Further, as shown in FIG. 14, in the above positioning device, the positioning data extraction module 134 to be matched may include:

The positioning data division unit 141 to be matched is used to divide the laser point cloud data into road surface laser point cloud data and / or road side laser point cloud data;

The positioning data extraction unit 142 to be matched is used for extracting laser point data of key points of roads and / or road objects on both sides of the road from the laser point cloud data of the road surface and / or the laser point cloud data of the road side.

The location data extraction module to be matched shown in FIG. 14 can be used to perform the method steps shown in FIG. 4a.

Further, as shown in FIG. 15, based on the structure shown in FIG. 14, the above positioning device may further include:

The road surface fitting module 151 is used to fit the road surface of the road according to the laser point cloud data of the road surface;

The data correction module 152 is used to correct the height value of the road surface laser point cloud data and / or the road side laser point cloud data based on the fitted road surface to a height value relative to the road surface.

The device structure shown in FIG. 15 can be used to perform the method steps shown in FIG. 5.

Further, as shown in FIG. 16, based on the structure shown in FIG. 14 or FIG. 15, the above-mentioned road object may be a ground mark on the road, and the location data extraction module 134 to be matched may include:

The road surface data dividing unit 161 is used to divide the road surface laser point cloud data into a grid according to a preset grid size;

The road surface data obtaining unit 162 is used to obtain the key point of a ground mark based on the laser point data of the ground mark in the grid if the road surface laser point cloud data in a grid contains the laser mark data of the ground mark Laser point data.

The location data extraction module to be matched shown in FIG. 16 can be used to perform the method steps shown in FIG. 6.

Further, as shown in FIG. 17, based on the structure shown in FIG. 14 or FIG. 15, the above-mentioned road object may be a road edge, and then the positioning data extraction module 134 to be matched may include:

The road side data dividing unit 171 is used to divide the road side laser point cloud data into a grid according to a preset grid size;

The road edge data sorting unit 172 is used to, if the road side laser point cloud data in a grid contains road edge laser point data, then the road edge laser point data according to the height value in the laser point data from small to large Order

The road edge data deleting unit 173 is configured to delete the laser point after the sort and the subsequent laser points of the two adjacent laser points if the difference between the height values of the two adjacent laser points after sorting is greater than the preset difference threshold Laser spot

The road edge data obtaining unit 174 is configured to obtain laser point data of a key point on the road edge based on the laser point data of the road edge retained in the grid.

The positioning data extraction module shown in FIG. 17 can be used to perform the method steps shown in FIG. 7a.

Further, as shown in FIG. 18, on the basis of the structure shown in FIG. 14 or FIG. 15, the above road object may be an upright object located on the side of the road, then the positioning data extraction module 134 to be matched may include:

The upright object data sorting unit 181 is used to, if the road-side laser point cloud data in a grid contains the laser point data of the road-side upright objects, the laser point data of the road-side upright objects according to the height in the laser point data The values are sorted in ascending order;

The upright object data deleting unit 182 is used to delete the laser points after the sort and the lasers after the sort if the height difference between the two adjacent laser points after sorting is greater than the preset difference threshold point;

The upright object data acquisition unit 183 is used to determine whether the lowest height value in the laser point data of the upright object is less than the preset first height threshold and whether the highest height value is greater than the preset second height threshold, and if so, then Based on the laser point data of the upright object retained in the grid, the laser point data of a key point of the upright object is obtained.

The to-be-matched positioning data extraction module shown in FIG. 18 can be used to perform the method steps shown in FIG. 8a.

Further, as shown in FIG. 19, on the basis of the structure shown in FIG. 14 or FIG. 15, the above road objects may include ground marks, road edges, and upright objects located on the side of the road, then the positioning data extraction module 134 to be matched may include:

The road surface and road side data dividing unit 191 is used to divide the road surface laser point cloud data and the road side laser point cloud data into a grid according to a preset grid size;

The pavement data unit 192 is used to obtain the laser of a key point of the ground mark based on the laser point data of the ground mark in the grid if the road surface laser point cloud data in a grid contains the laser mark data of the ground mark Point data

Road edge data unit 193, if the road side laser point cloud data in a grid contains road edge laser point data, then the road edge laser point data in the order of the height value in the laser point data from small to large Sorting, if the difference between the height values of the two adjacent laser spots after sorting is greater than the preset difference threshold, the laser spots after the sorting and the subsequent laser spots in the two adjacent laser spots are deleted, based on this network Retain the laser point data of the road edge in the grid to obtain the laser point data of a key point on the road edge;

The upright object data unit 194 is used to convert the laser point data of the road-side upright objects according to the height value in the laser point data if the road-side laser point cloud data in a grid contains the laser point data of the road-side upright objects Sort in ascending order. If the height difference between the two adjacent laser points after sorting is greater than the preset difference threshold, delete the laser point in the order and the laser points after it in the two adjacent laser points To determine whether the lowest height value in the laser point data of the retained upright objects is less than the preset first height threshold, and whether the highest height value is greater than the preset second height threshold, and if so, based on the uprights retained in the grid Laser point data of an object, to obtain laser point data of a key point of an upright object.

The to-be-matched positioning data extraction module shown in FIG. 19 can be used to perform the method steps shown in FIG. 9a.

Further, the sampling position of the vehicle at the initial time may be more than one location point selected from the positioning area where the GNSS positioning position of the vehicle is located at the initial time.

Further, as shown in FIG. 20, in the above positioning device, the sampling position acquisition module 135 may include:

The first sampling position extraction unit 201 is used to extract the first sampling position where the probability that the vehicle is at the corresponding sampling position is greater than the probability threshold from more than one sampling position corresponding to the vehicle at the previous moment;

The sampling position obtaining unit 203 is configured to forwardly simulate the motion state of the vehicle with respect to the first sampling position, and obtain more than one sampling position corresponding to the vehicle at this moment.

Further, the sampling position acquisition module shown in FIG. 20 may further include:

The first sampling position adding unit 202 is used to select a plurality of position points near the first sampling position as the additional first sampling position.

The sampling position acquisition module shown in FIG. 20 can be used to perform the method steps shown in FIG. 10.

Further, in any of the positioning devices shown above, the positioning data matching module 137 may be specifically used for,

Matching the positioning data to be matched in the converted coordinate system with the laser point data of the key points of the road objects of the same type in the standard positioning data, and based on the matching result, the probability that the vehicle is located at each sampling position is obtained.

Further, as shown in FIG. 21, the above positioning data matching module 137 may include:

The positioning data matching unit 211 is used to match the laser point data of the key point in the positioning data to be matched of the converted coordinate system with the laser point data of the closest key point in the standard positioning data, and calculate the position to be matched of the converted coordinate system The probability that the key points in the data are the same as the key points in the standard positioning data that are closest to each other;

The sampling position matching unit 212 is used to obtain the probability that the vehicle is located at the sampling position according to the probability that the key point in the positioning data to be matched of all the conversion coordinate systems corresponding to the same sampling position and the closest key point in the standard positioning data are the same position point Probability.

The positioning data matching module shown in FIG. 21 can be used to execute the contents of steps S111 and S121 to S122 in the methods shown in FIGS. 11 and 12.

Further, the above-mentioned high-precision positioning module 138 may be specifically used for,

The sampling position with the highest probability among the probability that the vehicle is located at each sampling position is determined as the high-precision positioning position of the vehicle;

or,

The invention provides a positioning device, based on the obtained GNSS positioning position of the vehicle, from the preset standard positioning data, to obtain the standard positioning data around the road where the vehicle is located; to obtain the laser around the road where the vehicle is output by the laser sensor Point cloud data, and from the laser point cloud data, extract the positioning data to be matched around the road where the vehicle is located; the above standard positioning data and the positioning data to be matched include those on the road where the vehicle is located and / or both sides of the road are easy to identify and have stable attributes Laser point data of key points of road objects; based on more than one sampling position corresponding to the vehicle at the previous moment, forward simulating the movement state of the vehicle, obtaining more than one sampling position corresponding to the vehicle at this moment, and based on the vehicle at this moment Corresponding to more than one sampling position, the positioning data to be matched is converted into the coordinate system corresponding to the standard positioning data; the positioning data to be matched in the converted coordinate system is matched with the standard positioning data, and based on the matching result, the vehicle is located in each Probability of sampling location; based on vehicle located in each Probability sampling position, high precision position location of the vehicle. Since the road object of the present invention is a road object that is easy to recognize and has stable attributes on the road and / or on both sides of the road, these road objects generally do not change due to changes in the environment or over time. Therefore, the road objects and / or Or the laser point data of the key points of the road objects that are easy to identify on both sides of the road and have stable attributes can be used as high-precision positioning matching objects to ensure the positioning success rate and accuracy. At the same time, the invention only extracts laser point data of key points of road objects for matching, so the amount of data is less, the calculation amount is greatly reduced, and the positioning efficiency is improved.

Further, by dividing the laser point cloud data into road surface laser point cloud data and / or road side laser point cloud data; then, from the road surface laser point cloud data and / or road side laser point cloud data, the road and And / or laser point data of key points of road objects on both sides of the road, so as to conveniently and quickly obtain laser point data of key points of road objects, that is, positioning data to be matched.

Further, by determining the road object as a ground mark on the road and dividing the road surface laser point cloud data into a grid according to a preset grid size; when judging the road surface laser point cloud data in a grid When the laser point data of the ground mark is included, the laser point data of a key point of the ground mark is obtained based on the laser point data of the ground mark in the grid, so that the laser point of the key point of the ground mark is obtained conveniently and quickly data.

Further, by determining the road object as a road edge, and dividing the road side laser point cloud data into a grid according to a preset grid size; when judging that the road side laser point cloud data in a grid contains the road edge The laser point data of the road, the laser point data on the edge of the road is sorted according to the order of the height values in the laser point data; if the difference between the height values of the two adjacent laser points after sorting is greater than the preset difference Threshold value, then delete the laser points that are sorted in the next two laser points and the subsequent laser points; finally, based on the laser point data of the road edge retained in the grid, obtain the laser of a key point on the road edge Point data, so that the laser point data of key points on the road edge can be obtained conveniently and quickly.

Further, by determining the road object as an upright object located on the road side, and dividing the road side laser point cloud data into a grid according to a preset grid size; when judging the road side laser point cloud in a grid When the data contains laser point data of upright objects on the road side, the laser point data of the upright objects on the road side are sorted in order of the height values in the laser point data; if the height values of two adjacent laser points after sorting If the difference is greater than the preset difference threshold, delete the laser point in the order and the laser points after it in the two adjacent laser points; finally, determine whether the lowest height value in the laser point data of the upright object retained Less than the preset first height threshold, whether the highest height value is greater than the preset second height threshold, and if so, based on the laser point data of the upright object retained in the grid, obtain the laser point of a key point of the upright object Data, so that the laser point data of the key points of upright objects on the road side can be obtained conveniently and quickly.

Further, by determining the road object as a ground mark on the road, a road edge, and an upright object on the side of the road, the laser point cloud data of the road surface and the laser point cloud data of the road side are divided into nets according to a preset grid size In the grid; and based on the laser point data of the ground marking in the grid, the laser point data of the road edge, and the laser point data of the upright objects on the side of the road, respectively obtain the laser point data of a key point of the corresponding road object, so as to achieve convenience , Quickly obtain laser point data of ground marks, key points of road edges and upright objects located on the side of the road.

Further, by extracting the first sampling position where the probability of the vehicle at the corresponding sampling position is greater than the probability threshold from more than one sampling position corresponding to the vehicle at the previous moment; and for the first sampling position, the vehicle motion state is simulated forward To obtain more than one sampling position corresponding to the vehicle at this moment, so as to realize convenient and rapid acquisition of multiple sampling positions of the vehicle used at this moment, and ensure the accuracy of the sampling position. In addition, by selecting a plurality of position points near the first sampling position as the additional first sampling position, the number of first sampling positions can be maintained at a certain level, while ensuring the accuracy of the sampling position.

Further, by matching the positioning data to be matched in the converted coordinate system with the laser point data of the key points of the road object of the same type in the standard positioning data, based on the matching result, the probability that the vehicle is located at each sampling position is obtained, so that In the process of matching the laser point data of key points, it is ensured that road objects of the same type are compared to improve the accuracy of the matching result, thereby improving the accuracy of the probability that the vehicle is located at each sampling position.

Example 10

The overall architecture of the positioning device has been described above. The function of the device can be implemented by means of an electronic device. As shown in FIG. 22, it is a schematic structural diagram of an electronic device according to an embodiment of the present invention, which specifically includes a memory 221 and a processor 222. .

The memory 221 is used to store programs.

In addition to the above programs, the memory 221 may be configured to store various other data to support operations on the electronic device. Examples of these data include instructions for any application or method for operating on the electronic device, contact data, phone book data, messages, pictures, videos, etc.

The memory 221 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable and removable Programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

The processor 222, coupled to the memory 221, is configured to execute a program in the memory 221, and the program executes any one of the positioning methods in FIGS. 3a to 12 when the program runs.

The above specific processing operations have been described in detail in the previous embodiments, and will not be repeated here.

Further, as shown in FIG. 22, the electronic device may further include: a communication component 223, a power component 224, an audio component 225, a display 226, and other components. Only some components are schematically shown in FIG. 22, and it does not mean that the electronic device includes only the components shown in FIG.

The communication component 223 is configured to facilitate wired or wireless communication between the electronic device and other devices. Electronic devices can access wireless networks based on communication standards, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 223 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 223 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

The power supply component 224 provides power for various components of the electronic device. The power supply component 224 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for electronic devices.

The audio component 225 is configured to output and / or input audio signals. For example, the audio component 225 includes a microphone (MIC). When the electronic device is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal. The received audio signal may be further stored in the memory 221 or transmitted via the communication component 223. In some embodiments, the audio component 225 further includes a speaker for outputting audio signals.

The display 226 includes a screen, which may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense the boundary of the touch or sliding action, but also detect the duration and pressure related to the touch or sliding operation.

Persons of ordinary skill in the art may understand that all or part of the steps of the foregoing method embodiments may be completed by a program instructing relevant hardware. The aforementioned program may be stored in a computer-readable storage medium. When the program is executed, the steps including the foregoing method embodiments are executed; and the foregoing storage medium includes various media that can store program codes, such as ROM, RAM, magnetic disk, or optical disk.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features can be equivalently replaced; and these modifications or replacements do not deviate from the essence of the corresponding technical solutions of the technical solutions of the embodiments of the present application range.

Claims

A positioning method, including:

Obtain the vehicle's GNSS positioning location;

Based on the vehicle's GNSS positioning position, standard positioning data around the road where the vehicle is located is obtained from preset standard positioning data, and the standard positioning data includes keys of road objects on the road and / or both sides of the road that are easy to identify and have stable attributes Point laser point data;

Obtain the laser point cloud data from the vehicle's laser sensor around the road where the vehicle is located;

From the laser point cloud data, extract to-be-matched positioning data around the road where the vehicle is located, the to-be-matched positioning data includes lasers of key points of road objects on the road where the vehicle is located and / or on both sides of the road that are easy to identify and have stable attributes Point data

Based on more than one sampling position corresponding to the vehicle at the previous moment, forward simulating the movement state of the vehicle to obtain more than one sampling position corresponding to the vehicle at this moment;

Convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data based on more than one sampling position corresponding to the vehicle at this moment;

Match the positioning data to be matched with the standard positioning data in the transformed coordinate system, and based on the matching result, obtain the probability that the vehicle is located at each sampling position;

Based on the probability that the vehicle is located at each sampling location, the highly precise location of the vehicle is obtained.
The method according to claim 1, wherein the extracting the positioning data to be matched around the road where the vehicle is located from the laser point cloud data includes:

Dividing the laser point cloud data into road surface laser point cloud data and / or road side laser point cloud data;

From the road surface laser point cloud data and / or road side laser point cloud data, extract laser point data of key points of roads and / or road objects on both sides of the road.
The method according to claim 2, wherein before the method extracts laser point data of key points, the method further comprises:

Fitting the road surface of the road according to the laser point cloud data of the road surface;

Based on the fitted road surface, the height value of the road surface laser point cloud data and / or road side laser point cloud data is corrected to a height value relative to the road surface.
The method according to claim 2 or 3, wherein the road object is a ground mark on a road, and then the laser point data of key points of the road object of the road is extracted from the road surface laser point cloud data This includes:

Divide the road surface laser point cloud data into a grid according to a preset grid size;

If the road surface laser point cloud data in a grid contains laser mark data of ground marks, then based on the laser point data of ground marks in the grid, the laser point data of a key point of the ground marks is acquired.
The method according to claim 2 or 3, wherein the road object is a road edge, and then the laser point data of key points of road objects on both sides of the road are extracted from the road side laser point cloud data include:

Divide the road side laser point cloud data into a grid according to a preset grid size;

If the road side laser point cloud data in a grid contains road edge laser point data, the road edge laser point data is sorted according to the order of the height values in the laser point data in ascending order;

If the difference between the height values of the two adjacent laser points after sorting is greater than the preset difference threshold, delete the laser points that are ranked after and after the laser points in the two adjacent laser points;

Based on the laser point data of the road edge retained in the grid, the laser point data of a key point on the road edge is obtained.
The method according to claim 2 or 3, wherein the road object is an upright object located on the side of the road, then the key points of the road objects on both sides of the road are extracted from the road-side laser point cloud data The laser spot data specifically includes:

Divide the road side laser point cloud data into a grid according to a preset grid size;

If the road-side laser point cloud data in a grid includes the laser point data of the road-side upright objects, the laser point data of the road-side upright objects is arranged in the order of the height value in the laser point data from small to large Sorting

If the height difference between the two adjacent laser points after sorting is greater than the preset difference threshold, delete the laser point after the sort and the laser points after it from the two adjacent laser points;

Determine whether the lowest height value in the laser point data of the upright object retained is less than the preset first height threshold, and whether the highest height value is greater than the preset second height threshold, and if so, based on the upright object retained in the grid To obtain the laser point data of a key point of an upright object.
The method according to claim 2 or 3, characterized in that the road objects include ground markings, road edges and upright objects located on the side of the road, then from the road surface laser point cloud data and road side laser point cloud data In the laser point data extraction of key points of roads and road objects on both sides of the road specifically include:

Divide the road surface laser point cloud data and road side laser point cloud data into grids according to a preset grid size;

If the road surface laser point cloud data in a grid includes laser mark data of ground marks, then based on the laser mark data of ground marks in the grid, obtain laser point data of a key point of the ground marks;

If the road side laser point cloud data in a grid contains road edge laser point data, the road edge laser point data is sorted according to the order of the height values in the laser point data from small to large. If the difference between the height values of the two adjacent laser points is greater than the preset difference threshold, the laser points that are sorted in the next two laser points and the subsequent laser points are deleted, based on the road edges retained in the grid Laser point data to obtain laser point data of a key point on the road edge;

If the road-side laser point cloud data in a grid includes the laser point data of the road-side upright objects, the laser point data of the road-side upright objects is arranged in the order of the height value in the laser point data from small to large Sorting, if the difference between the heights of two adjacent laser points after sorting is greater than the preset difference threshold, delete the laser points that are sorted in the next two laser points and the subsequent laser points, and judge the remaining upright objects Whether the lowest height value in the laser point data of is less than the preset first height threshold, and whether the highest height value is greater than the preset second height threshold, and if so, based on the laser point data of upright objects retained in the grid, Acquire laser point data of a key point of an upright object.
The method according to claim 1, wherein matching the positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, obtaining the probability that the vehicle is located at each sampling position includes:

Matching the positioning data to be matched in the converted coordinate system with the laser point data of the key points of the road object of the same type in the standard positioning data, and based on the matching result, the probability that the vehicle is located at each sampling position is obtained.
The method according to claim 8, wherein the matching positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, obtaining the probability that the vehicle is located at each sampling position includes:

Matching the laser point data of the key point in the positioning data to be matched of the converted coordinate system with the laser point data of the closest key point in the standard positioning data, and calculating the positioning data to be matched in the converted coordinate system The probability that the key point of the key point and the closest key point in the standard positioning data are the same position point;

The probability that the vehicle is located at the sampling position is obtained according to the probability that the key point in the positioning data to be matched of all the converted coordinate systems corresponding to the same sampling position and the closest key point in the standard positioning data are the same position point.
The method according to claim 1, wherein the obtaining the high-precision positioning position of the vehicle based on the probability that the vehicle is located at each sampling position includes:

Determining the sampling position with the highest probability among the probability that the vehicle is located at each sampling position as the high-precision positioning position of the vehicle;

or,

Taking the probability that the vehicle is located at each sampling position as a weight, weighting calculation is performed on each corresponding sampling position, and the obtained weighted position is determined as the high-precision positioning position of the vehicle.
A positioning device, including:

GNSS positioning position acquisition module, used to obtain the GNSS positioning position of the vehicle;

The standard positioning data acquisition module is used to obtain standard positioning data around the road on which the vehicle is located based on the GNSS positioning position of the vehicle, the standard positioning data including the road and / or both sides of the road is easy to identify And the laser point data of key points of road objects with stable attributes;

The laser point cloud data acquisition module is used to obtain laser point cloud data around the road where the vehicle is output by the laser sensor of the vehicle;

A positioning data extraction module to be matched is used to extract positioning data to be matched around the road where the vehicle is located from the laser point cloud data, the positioning data to be matched includes the road on which the vehicle is located and / or both sides of the road are easy to identify and have attributes Laser point data of key points of stable road objects;

The sampling position acquisition module is used to forward simulate the movement state of the vehicle based on more than one sampling position corresponding to the vehicle at the previous moment, and acquire more than one sampling position corresponding to the vehicle at the moment;

The positioning data conversion module is used to convert the positioning data to be matched into the coordinate system corresponding to the standard positioning data based on more than one sampling position corresponding to the vehicle at this moment;

The positioning data matching module is used to match the positioning data to be matched with the standard positioning data in the converted coordinate system, and based on the matching result, obtain the probability that the vehicle is located at each sampling position;

The high-precision positioning module is used to obtain the high-precision positioning position of the vehicle based on the probability that the vehicle is located at each sampling position.
An electronic device, including:

Memory for storing programs;

A processor, coupled to the memory, is used to execute the program, and when the program runs, the positioning method according to any one of claims 1-10 is executed.